Application and evaluation of classification trees for screening unwanted plants

Abstract Risk assessment systems for introduced species are being developed and applied globally, but methods for rigorously evaluating them are still in their infancy. We explore classification and regression tree models as an alternative to the current Australian Weed Risk Assessment system, and demonstrate how the performance of screening tests for unwanted alien species may be quantitatively compared using receiver operating characteristic (ROC) curve analysis. The optimal classification tree model for predicting weediness included just four out of a possible 44 attributes of introduced plants examined, namely: (i) intentional human dispersal of propagules; (ii) evidence of naturalization beyond native range; (iii) evidence of being a weed elsewhere; and (iv) a high level of domestication. Intentional human dispersal of propagules in combination with evidence of naturalization beyond a plants native range led to the strongest prediction of weediness. A high level of domestication in combination with no evidence of naturalization mitigated the likelihood of an introduced plant becoming a weed resulting from intentional human dispersal of propagules. Unlikely intentional human dispersal of propagules combined with no evidence of being a weed elsewhere led to the lowest predicted probability of weediness. The failure to include intrinsic plant attributes in the model suggests that either these attributes are not useful general predictors of weediness, or data and analysis were inadequate to elucidate the underlying relationship(s). This concurs with the historical pessimism that we will ever be able to accurately predict invasive plants. Given the apparent importance of propagule pressure (the number of individuals of an species released), future attempts at evaluating screening model performance for identifying unwanted plants need to account for propagule pressure when collating and/or analysing datasets. The classification tree had a cross‐validated sensitivity of 93.6% and specificity of 36.7%. Based on the area under the ROC curve, the performance of the classification tree in correctly classifying plants as weeds or non‐weeds was slightly inferior (Area under ROC curve = 0.83 ± 0.021 (±SE)) to that of the current risk assessment system in use (Area under ROC curve = 0.89 ± 0.018 (±SE)), although requires many fewer questions to be answered.     

Drought impact on water use efficiency and intra‐annual density fluctuations in Erica arborea on Elba (Italy)

Erica arborea (L) is a widespread Mediterranean species, able to cope with water stress and colonize semiarid environments. The eco‐physiological plasticity of this species was evaluated by studying plants growing at two sites with different soil moistures on the island of Elba (Italy), through dendrochronological, wood‐anatomical analyses and stable isotopes measurements. Intra‐annual density fluctuations (IADFs) were abundant in tree rings, and were identified as the key parameter to understand site‐specific plant responses to water stress. Our findings showed that the formation of IADFs is mainly related to the high temperature, precipitation patterns and probably to soil water availability, which differs at the selected study sites. The recorded increase in the ¹³C‐derived intrinsic water use efficiency at the IADFs level was linked to reduced water loss rather than to increasing C assimilation. The variation in vessel size and the different absolute values of δ¹⁸O among trees growing at the two study sites underlined possible differences in stomatal control of water loss and possible differences in sources of water uptake. This approach not only helped monitor seasonal environmental differences through tree‐ring width, but also added valuable information on E. arborea responses to drought and their ecological implications for Mediterranean vegetation dynamics.     

Interactive effects of drought and N fertilization on the spatial distribution of methane assimilation in grassland soils

Soil methanotrophic bacteria constitute the only globally relevant biological sink for atmospheric methane (CH₄). Nitrogen (N) fertilizers as well as soil moisture regime affect the activity of these organisms, but the mechanisms involved are not well understood to date. In particular, virtually nothing is known about the spatial distribution of soil methanotrophs within soil structure and how this regulates CH₄ fluxes at the ecosystem scale. We studied the spatial distribution of CH₄ assimilation and its response to a factorial drought × N fertilizer treatment in a 3‐year experiment replicated in two grasslands differing in management intensity. Intact soil cores were labelled with ¹⁴CH₄ and methanotrophic activity mapped at a resolution of ～100 μm using an autoradiographic technique. Under drought, the main zone of CH₄ assimilation shifted down the soil profile. Ammonium nitrate (NH₄NO₃) and cattle urine reduced CH₄ assimilation in the top soil, but only when applied under drought, presumably because NH₄⁺ from fertilizers was not removed by plant uptake and nitrification under these conditions. Ecosystem‐level CH₄ fluxes measured in the field did show no or only very small inhibitory effects, suggesting that deeper soil layers fully compensated for the reduction in top soil CH₄ assimilation. Our results indicate that the ecosystem‐level CH₄ sink cannot be inferred from measurements of soil samples that do not reflect the spatial organization of soils (e.g. stratification of organisms, processes, and mechanisms). The autoradiographic technique we have developed is suited to study methanotrophic activity in a relevant spatial context and does not rely on the genetic identity of the soil bacterial communities involved, thus ideally complementing DNA‐based approaches.     

Aminopeptidase and phosphatase activities in basins of Lake Hiidenvesi dominated by cyanobacteria and in laboratory grown Anabaena

1. Extracellular enzyme activities were examined in freshwater basins representing a transition from hypertrophy to mesotrophy and in axenic cyanobacterial cultures to evaluate the ecological role of extracellular enzyme activities of cyanobacteria.
2. Aminopeptidase activity was related to the trophic status of the lake basins. The activity was highest in the most eutrophic basin and decreased in the less nutrient-rich basins. Cyanobacteria were the most important autotrophic organisms and aminopeptidase activity was positively associated with cyanobacterial biomass.
3. In an axenic Anabaena batch culture, nitrogenase activity was several orders of magnitude higher than leucine aminopeptidase activity. Nitrate did not have an effect on aminopeptidase activity or growth, but significantly reduced the rate of nitrogen fixation. A high phosphorus concentration at the beginning of the Anabaena batch-culture experiment resulted in reduced phosphatase activity.
4. In Lake Hiidenvesi, aminopeptidase activity probably originated mostly from attached bacteria and less so from cyanobacteria.     

Benthic–pelagic coupling in the population dynamics of the harmful cyanobacterium Microcystis

1. In eutrophic lakes, large amounts of the cyanobacterium Microcystis may overwinter in the sediment and re‐inoculate the water column in spring. 2. We monitored changes in pelagic and benthic populations of Microcystis in Lake Volkerak, The Netherlands. In addition, sedimentation rates and the rate of recruitment from the sediment were measured using traps. These data were used to model the coupling between the benthic and pelagic populations and to calculate the contribution of overwintering benthic and pelagic populations to the magnitude of the pelagic summer bloom. 3. Changes in the benthic Microcystis population showed a time lag of 3–14 weeks compared with the pelagic population. This time lag increased with lake depth. The largest amount of benthic Microcystis was found in the deepest parts of the lake. These observations suggest horizontal transport of sedimented Microcystis from shallow to deep parts of the lake. 4. Recruitment from and sedimentation to the sediment occurred throughout the year, with highest recruitment and sedimentation rates during summer. Model simulations indicate that the absence of benthic recruitment would reduce the summer bloom by 50%. 5. In spring, the total pelagic population was three to six times smaller than the total benthic population. Yet, model simulations predict that the absence of this small overwintering pelagic population would reduce the summer bloom by more than 64%. 6. Reduction of the overwintering pelagic populations, for instance by flushing, may be a useful management strategy to suppress or at least delay summer blooms of Microcystis.     

Artificial mixing prevents nuisance blooms of the cyanobacterium Microcystis in Lake Nieuwe Meer, the Netherlands

1. Artificial mixing in the hypertrophic Lake Nieuwe Meer was successful in preventing blooms of the cyanobacterium Microcystis. During the 2 years of artificial, deep mixing the number of colonies of Microcystis per litre and also per m² was lower than in the two preceding control years. Hardly any nuisance scums of Microcystis occurred in the lake.
2. The phytoplankton shifted from a cyanobacteria-dominated community in summer to a mixed community of flagellates, green algae and diatoms. Reduced sedimentation losses in the mixed lake, probably in combination with a lower pH, favoured non-buoyant algae, while the entrainment of cyanobacteria in the turbulent flow nullified their advantage of buoyancy.
3. The chlorophyll concentrations were much lower in the mixed lake, but the euphotic depth did not show clear differences between the years. The chlorophyll content integrated through depth (m^–2) increased in the artificially mixed lake.
4. The deep lake normally stratified in summer, but artificial mixing of the lake in 1993 resulted in a homogeneous temperature and oxygen distribution with depth. In spring 1994, the mixing was applied intermittently with a reduction of 75% of the energy costs, while the mixing was still sufficient to prevent stratification.
5. Determination of the buoyancy state of the colonies on a sunny and calm day showed that the buoyancy loss was low close to the bubble plumes, and high at some distance from these plumes. This suggests that Microcystis could escape the mixing at some distance from the plumes, and could synthesize more carbohydrates during its stay in the upper illuminated layer of the lake than the deep mixed colonies close to the bubble plumes. Determination of the buoyancy state appeared to be a good and simple method to investigate the extent of entrainment of colonies in the turbulent flow.     

Intracellular pH stability in the aquatic resurrection plant Chamaegigas intrepidus in the extreme environmental conditions that characterize its natural habitat

Chamaegigas intrepidus Dinter (syn. Lindernia intrepidus (Dinter) Oberm.) is a poikilohydric aquatic plant that lives in rock pools on granitic outcrops in Central Namibia. The pools are only filled intermittently during the summer rains, and the plants can pass through 15–20 rehydration/dehydration cycles during a single wet season. Rehydrated plants also have to cope with substantial diurnal fluctuations in the pool pH as a result of photosynthetic CO₂ uptake. We have used in vivo ³¹P NMR spectroscopy to investigate the effect of external pH and dehydration (low water potential) on intracellular pH in the roots and submerged leaves of C. intrepidus . Increasing the external pH from 6 to 10 had no effect on the steady state cytoplasmic and vacuolar pH values of submerged leaves, but caused a slight alkalinization of the root cytoplasm. Similarly dehydration with PEG-600 at either pH 6 or pH 10 had no effect on the cytoplasmic pH of the leaves, but it did cause a small alkalinization of the leaf vacuoles at pH 10. These results imply an unusually effective regulation of intracellular pH, consistent with the adaptation of C. intrepidus to the extreme environmental conditions of its habitat. The NMR analysis also showed that dehydration had no effect on the inorganic phosphate and phosphocholine pools, and this was taken to indicate that the cell membranes were well protected from the effects of the low water potential.     

Trophic level modulates carabid beetle responses to habitat and landscape structure: a pan‐European study

1. Anthropogenic pressures have produced heterogeneous landscapes expected to influence diversity differently across trophic levels and spatial scales. 2. We tested how activity density and species richness of carabid trophic groups responded to local habitat and landscape structure (forest percentage cover and habitat richness) in 48 landscape parcels (1 km²) across eight European countries. 3. Local habitat affected activity density, but not species richness, of both trophic groups. Activity densities were greater in rotational cropping compared with other habitats; phytophage densities were also greater in grassland than forest habitats. 4. Controlling for country and habitat effects, we found general trophic group responses to landscape structure. Activity densities of phytophages were positively correlated, and zoophages uncorrelated, with increasing habitat richness. This differential functional group response to landscape structure was consistent across Europe, indicated by a lack of a country × habitat richness interaction. Species richness was unaffected by landscape structure. 5. Phytophage sensitivity to landscape structure may arise from relative dependency on seed from ruderal plants. This trophic adaptation, rare in Carabidae, leads to lower phytophage numbers, increasing vulnerability to demographic and stochastic processes that the greater abundance, species richness, and broader diet of the zoophage group may insure against.